Paul A. Bristow wrote:
-----Original Message----- From: boost-users-bounces@lists.boost.org [mailto:boost-users- bounces@lists.boost.org] On Behalf Of Neal Becker Sent: Friday, November 05, 2010 10:38 AM To: boost-users@lists.boost.org Subject: Re: [Boost-users] Statistical distributions in Boost.Math
Paul A. Bristow wrote:
Recently I have been working on a few more Statistical Distributions for Boost.Math.
If you have suggestions or requests, please tell me.
I am needing cdf of doubly-noncentric F distribution, and having trouble finding code (execpt in mathematica). I have found a few papers, but haven't found time to review them.
Hmm - looks non-trivial :-(
http://mathworld.wolfram.com/NoncentralF-Distribution.html gives the complex details.
As I recall, hypergeometric functions caused so much numerical trouble that they were withdrawn from TR1 (and not implemented in Boost.math)? Or is that something different?
A quick Google references some promising looking articles, but I don't have access to most of them. If you can provide them as pdf, I'll think about it more carefully.
Paul
I've found some references, but I'm no mathematician. The one I've toyed with, because it was easy to understand, is: http://www.ifpress.com/pdfs/comm22-897.pdf This is a numerical integration. This sounds interesting, but I haven't looked at it yet: http://onlinelibrary.wiley.com/doi/10.1111/j.1467-842X.1965.tb00036.x/abstra... Some others: http://onlinelibrary.wiley.com/doi/10.1111/j.1467-842X.1969.tb00102.x/abstra... http://fmwww.bc.edu/cef99/papers/Paolella.pdf